Hello @ChetanKarwa,

great to see you managed to build stdlib and get your new function working.

However, I am not totally convinced by the purpose of this function. The more common usage of a sort function would be to sort a list of strings (like the one in development in #311), say to output a sorted list of labels.

Do you have some usage cases that motivate a character sort function?

I implemented this because I've seen other languages supporting the sorting of a character array.
But they support a general sort function they don't have a separate sort function for the character array and that function is slow [time complexity - O(N*logN)]. So I thought of implementing something different for Fortran stdlib and so the one implemented in this PR is O(N).
Now giving a second thought to it currently I really can't figure out any use case of this function.

@wclodius2 Is your implementation related to this PR?

No, it isn't. My sort is of an array of elements of the string_type and not of the characters of a string_type scalar, or of a list of elements of the string_type. FWIW I also view sorting of the characters as of very limited use, and sorting a list rather than an array would require aa drastic rewrite of my code.

I renamed this PR to avoid confusion with #408

What could be done is to provide the O(n) sorting algorithm of @ChetanKarwa as a specialization in #408 when the input is an array of character(len=1), but I don't expect this to be used very often.

The only usage I can think of currently is to find the median letter (and maybe the mode) in a character sequence.

Also, consider that this implementation is likely to fail for non-ASCII input:

print *, sort("Ångström")
! Fortran runtime error: Index '195' of dimension 1 of array 'count' above upper bound of 128